CN112309366A - Speech synthesis method, speech synthesis device, storage medium and electronic equipment - Google Patents
Speech synthesis method, speech synthesis device, storage medium and electronic equipment Download PDFInfo
- Publication number
- CN112309366A CN112309366A CN202011211084.5A CN202011211084A CN112309366A CN 112309366 A CN112309366 A CN 112309366A CN 202011211084 A CN202011211084 A CN 202011211084A CN 112309366 A CN112309366 A CN 112309366A
- Authority
- CN
- China
- Prior art keywords
- text
- sample
- accent
- vector
- phoneme
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000015572 biosynthetic process Effects 0.000 title claims abstract description 143
- 238000003786 synthesis reaction Methods 0.000 title claims abstract description 143
- 238000001308 synthesis method Methods 0.000 title abstract description 18
- 238000012549 training Methods 0.000 claims abstract description 48
- 239000013598 vector Substances 0.000 claims description 255
- 238000000034 method Methods 0.000 claims description 82
- 238000001228 spectrum Methods 0.000 claims description 67
- 230000008569 process Effects 0.000 claims description 29
- 238000012545 processing Methods 0.000 claims description 29
- 238000004590 computer program Methods 0.000 claims description 19
- 238000006243 chemical reaction Methods 0.000 claims description 16
- 230000006870 function Effects 0.000 description 87
- 238000010586 diagram Methods 0.000 description 11
- 238000004364 calculation method Methods 0.000 description 10
- 238000004891 communication Methods 0.000 description 7
- 230000003287 optical effect Effects 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 5
- 238000013459 approach Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 230000011218 segmentation Effects 0.000 description 3
- 238000002372 labelling Methods 0.000 description 2
- 239000013307 optical fiber Substances 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- 239000004065 semiconductor Substances 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000033764 rhythmic process Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000035807 sensation Effects 0.000 description 1
- 230000002194 synthesizing effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L13/00—Speech synthesis; Text to speech systems
- G10L13/02—Methods for producing synthetic speech; Speech synthesisers
- G10L13/033—Voice editing, e.g. manipulating the voice of the synthesiser
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L13/00—Speech synthesis; Text to speech systems
- G10L13/02—Methods for producing synthetic speech; Speech synthesisers
- G10L13/04—Details of speech synthesis systems, e.g. synthesiser structure or memory management
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L13/00—Speech synthesis; Text to speech systems
- G10L13/08—Text analysis or generation of parameters for speech synthesis out of text, e.g. grapheme to phoneme translation, prosody generation or stress or intonation determination
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L13/00—Speech synthesis; Text to speech systems
- G10L13/02—Methods for producing synthetic speech; Speech synthesisers
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L13/00—Speech synthesis; Text to speech systems
- G10L13/02—Methods for producing synthetic speech; Speech synthesisers
- G10L13/04—Details of speech synthesis systems, e.g. synthesiser structure or memory management
- G10L13/047—Architecture of speech synthesisers
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Machine Translation (AREA)
- Document Processing Apparatus (AREA)
Abstract
The present disclosure relates to a voice synthesis method, apparatus, storage medium, and electronic device to include accents in synthesized voice and control the timbre of the synthesized voice. The speech synthesis method comprises the following steps: acquiring a text to be synthesized with a target tone label, wherein the target tone label is a tone label of a first speaker or a tone label of a second speaker; inputting a text to be synthesized into a voice synthesis model to obtain audio information corresponding to the text to be synthesized, wherein training samples of the voice synthesis model comprise a first sample text marked with accent words, a first sample audio of a first speaker to the first sample text, a second sample text not marked with the accent words and a second sample audio of a second speaker to the second sample text, and the voice synthesis model is used for determining the accent words in the text to be synthesized and generating the audio information corresponding to the text to be synthesized according to the accent words and a target tone label.
Description
Technical Field
The present disclosure relates to the field of speech synthesis technologies, and in particular, to a speech synthesis method, apparatus, storage medium, and electronic device.
Background
Speech synthesis, also known as Text To Speech (TTS), is a technology that can convert any input Text into corresponding Speech. Conventional speech synthesis systems typically include two modules, a front end and a back end. The front-end module mainly analyzes the input text and extracts the linguistic information needed by the rear-end module. And the back-end module generates a voice waveform by a certain method according to the front-end analysis result.
However, the speech synthesis method in the related art generally does not consider accents in the synthesized speech, and thus the synthesized speech is not accented, has a flat pronunciation, and lacks expressive power. Or, the speech synthesis method in the related art generally randomly selects words in the input text for accent addition, which causes an error in pronunciation of accents in the synthesized speech and fails to obtain a better speech synthesis result including accents.
Disclosure of Invention
This summary is provided to introduce a selection of concepts in a simplified form that are further described below in the detailed description. This summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.
In a first aspect, the present disclosure provides a method of speech synthesis, the method comprising:
acquiring a text to be synthesized with a target tone label, wherein the target tone label is a tone label of a first speaker or a tone label of a second speaker;
inputting the text to be synthesized into a speech synthesis model to obtain audio information corresponding to the text to be synthesized, wherein training samples of the speech synthesis model comprise a first sample text marked with accent words, a first sample audio of the first speaker to the first sample text, a second sample text not marked with accent words, and a second sample audio of the second speaker to the second sample text;
the voice synthesis model is used for determining accent words in the text to be synthesized and generating audio information corresponding to the text to be synthesized according to the accent words and the target tone label.
In a second aspect, the present disclosure provides a speech synthesis apparatus, the apparatus comprising:
the acquisition module is used for acquiring a text to be synthesized with a target tone label, wherein the target tone label is the tone label of the first speaker or the tone label of the second speaker;
a synthesis module, configured to input the text to be synthesized into a speech synthesis model to obtain audio information corresponding to the text to be synthesized, where a training sample of the speech synthesis model includes a first sample text labeled with an accent word and a first sample audio of the first sample text by the first speaker, and a second sample text not labeled with an accent word and a second sample audio of the second sample text by the second speaker;
the voice synthesis model is used for determining accent words in the text to be synthesized and generating audio information corresponding to the text to be synthesized according to the accent words and the target tone label.
In a third aspect, the present disclosure provides a computer readable medium having stored thereon a computer program which, when executed by a processing apparatus, performs the steps of the method of the first aspect.
In a fourth aspect, the present disclosure provides an electronic device comprising:
a storage device having a computer program stored thereon;
processing means for executing the computer program in the storage means to carry out the steps of the method of the first aspect.
Through the technical scheme, the synthesized voice can comprise accent pronunciation, and the voice synthesis model can train the first sample audio of the first sample text through the first sample text marked with accent words and the first speaker, so that compared with a mode of randomly adding accent pronunciation in the related technology, the accuracy of the accent pronunciation in the generated audio information can be ensured to a certain extent. In addition, the accent pronunciation habit of the first speaker can be transferred to the synthesized voice corresponding to the second speaker through the method, namely the synthesized voice with the accent pronunciation habit as the first speaker and the timbre as the second speaker can be obtained, the accented voice with the appointed timbre is synthesized, and under the condition that the accented audio of the second speaker is needed to be used, the accented audio does not need to be manually recorded by the second speaker, so that the labor and time consumed for obtaining the accented audio are reduced.
Additional features and advantages of the disclosure will be set forth in the detailed description which follows.
Drawings
The above and other features, advantages and aspects of various embodiments of the present disclosure will become more apparent by referring to the following detailed description when taken in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numbers refer to the same or similar elements. It should be understood that the drawings are schematic and that elements and features are not necessarily drawn to scale. In the drawings:
FIG. 1 is a flow diagram illustrating a method of speech synthesis according to an exemplary embodiment of the present disclosure;
FIG. 2 is a schematic diagram illustrating a speech synthesis model in a method of speech synthesis according to an exemplary embodiment of the present disclosure;
FIG. 3 is a schematic diagram illustrating a speech synthesis model in a method of speech synthesis according to another exemplary embodiment of the present disclosure;
FIG. 4 is a schematic diagram illustrating a speech synthesis model in a method of speech synthesis according to another exemplary embodiment of the present disclosure;
FIG. 5 is a block diagram illustrating a speech synthesis apparatus according to an exemplary embodiment of the present disclosure;
fig. 6 is a block diagram illustrating an electronic device according to an exemplary embodiment of the present disclosure.
Detailed Description
Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present disclosure are shown in the drawings, it is to be understood that the present disclosure may be embodied in various forms and should not be construed as limited to the embodiments set forth herein, but rather are provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the disclosure are for illustration purposes only and are not intended to limit the scope of the disclosure.
It should be understood that the various steps recited in the method embodiments of the present disclosure may be performed in a different order, and/or performed in parallel. Moreover, method embodiments may include additional steps and/or omit performing the illustrated steps. The scope of the present disclosure is not limited in this respect.
The term "include" and variations thereof as used herein are open-ended, i.e., "including but not limited to". The term "based on" is "based, at least in part, on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Relevant definitions for other terms will be given in the following description. It should be noted that the terms "first", "second", and the like in the present disclosure are only used for distinguishing different devices, modules or units, and are not used for limiting the order or interdependence relationship of the functions performed by the devices, modules or units. It is further noted that references to "a", "an", and "the" modifications in the present disclosure are intended to be illustrative rather than limiting, and that those skilled in the art will recognize that "one or more" may be used unless the context clearly dictates otherwise.
The names of messages or information exchanged between devices in the embodiments of the present disclosure are for illustrative purposes only, and are not intended to limit the scope of the messages or information.
As mentioned in the background, the speech synthesis method in the related art usually does not consider accents in the synthesized speech, resulting in no accents, flat pronunciation, and lack of expressiveness in the synthesized speech. Or, the speech synthesis method in the related art generally randomly selects words in the input text for accent addition, which causes an error in pronunciation of accents in the synthesized speech and fails to obtain a better speech synthesis result including accents.
In view of the above, the present disclosure provides a speech synthesis method, apparatus, storage medium and electronic device, which are capable of controlling the timbre of the synthesized speech and synthesizing accented speech with a specified timbre by a new speech synthesis method, so as to improve the accuracy of accents in the synthesized speech.
Fig. 1 is a flow chart illustrating a method of speech synthesis according to an exemplary embodiment of the present disclosure. Referring to fig. 1, the speech synthesis method includes:
the voice synthesis model is used for determining accent words in the text to be synthesized and generating audio information corresponding to the text to be synthesized according to the accent words and the target tone label.
Through the method, the synthesized voice can comprise the accent pronunciation, and the voice synthesis model can train the first sample audio of the first sample text through the first sample text marked with the accent words and the first speaker, so that compared with a method of randomly adding the accent pronunciation in the related technology, the accuracy of the accent pronunciation in the generated audio information can be ensured to a certain extent. In addition, the accent pronunciation habit of the first speaker can be transferred to the synthesized voice corresponding to the second speaker through the method, namely the synthesized voice with the accent pronunciation habit as the first speaker and the timbre as the second speaker can be obtained, the accented voice with the appointed timbre is synthesized, and under the condition that the accented audio of the second speaker is needed to be used, the accented audio does not need to be manually recorded by the second speaker, so that the labor and time consumed for obtaining the accented audio are reduced.
In order to make the speech synthesis method provided by the present disclosure more understandable to those skilled in the art, the above steps are exemplified in detail below.
First, a training process of the speech synthesis model is explained.
For example, a plurality of first sample texts and first sample audios of a first speaker for the plurality of first sample texts for training, and a plurality of second sample texts and second sample audios of a second speaker for the plurality of second sample texts may be obtained in advance, wherein each first sample text is marked with an accent word, each second sample text is not marked with an accent, that is, each first sample text is marked with a word requiring accent pronunciation, and each second sample text is not marked with a word requiring accent pronunciation.
In a possible manner, the accented words noted in the first sample text may be determined by: the method comprises the steps of firstly obtaining a plurality of first sample texts, wherein each first sample text comprises an accent word marked with an initial accent mark, then adding a target accent mark to each accent word marked with the initial accent mark if the accent word is marked as an accent word in each first sample text, adding a target accent mark to each accent word if the accent word is marked as an accent word in at least two first sample texts, adding the target accent mark to each accent word if the fundamental frequency of the accent word is larger than a preset fundamental frequency threshold and the energy of the accent word is larger than a preset energy threshold, and finally determining the accent word added with the target accent mark in each first sample text as the accent word in the first sample text.
The plurality of first samples may be sample texts including the same content and subjected to initial accent marking by different users, or the plurality of first samples may be a plurality of texts including different content and subjected to initial accent marking by different users, and so on, which are not limited by the embodiment of the present disclosure. It should be understood that, in order to improve the result accuracy, it is preferable that the plurality of first sample texts are a plurality of texts including different contents and the texts including the same content are initially accented by different users.
For example, the time boundary information of each word and each prosodic phrase in the first sample text may be obtained by first obtaining the time boundary information of each word in the first sample text through the auto-alignment model. Then, the multiple users can mark the accent words at the prosodic phrase level according to the aligned first sample audio and first sample text, and by combining the listening sensation, the oscillogram, the frequency spectrum, and the semantic information obtained from the first sample text, to obtain multiple first sample texts with the accent words marked thereon. Wherein the prosodic phrases are intermediate rhythm chunks between the prosodic words and the intonation phrases. Prosodic words are a set of closely related syllables that are often pronounced together in the actual stream of speech. The intonation phrases are formed by connecting several prosodic phrases according to a certain intonation pattern, and generally correspond to sentences in syntax. In the embodiment of the present disclosure, the initial accent marks in the first sample text may correspond to prosodic phrases, so as to obtain prosodic phrase-level initial accent marks, so that the accent pronunciation is more in line with conventional pronunciation habits.
Alternatively, in other possible cases, the initial accent marks in the first sample text may correspond to single words, so as to obtain word-level accents or single-word-level accents, and so on, which may be selected according to requirements in specific implementations.
After obtaining the plurality of first sample texts with the initial accent marks, the initial accent marks in the plurality of first sample texts may be integrated. Specifically, for each accent word marked with an initial accent mark, if the accent word is marked as an accent word in each first sample text, it indicates that the accent marking result is more accurate, and therefore, a target accent mark may be added to the accent word. If the accent word is marked as an accent word in at least two first sample texts, the fact that the accent word is not marked as an accent exists in other first sample texts is explained, and therefore a certain deviation may exist in the accent marking result. In this case, in order to improve the accuracy of the result, further judgment may be performed. Specifically, considering that the fundamental frequency of the accent pronunciation in the audio is higher than the fundamental frequency of the non-accent pronunciation, and the energy of the accent pronunciation in the audio is higher than the energy of the non-accent pronunciation, the target accent mark may be added to the accent word under the condition that the fundamental frequency of the accent word is greater than a preset fundamental frequency threshold and the energy of the accent word is greater than a preset energy threshold. The preset fundamental frequency threshold and the preset energy threshold may be set according to actual conditions, which is not limited in the embodiments of the present disclosure.
It should be understood that, in other possible cases, if a stressed word is not labeled as stressed in all other first sample texts, the probability that the stressed word is stressed is low, so that the stressed word may not be added with a target stressed mark.
By the method, the first sample text marked with the initial accent marks can be subjected to accent mark screening, and the first sample text added with the target accent marks is obtained, so that the accent words added with the target accent marks can be determined as the accent words in the first sample text aiming at each first sample text, and the accent mark information in the first sample text is more accurate.
After the first sample texts marked with the accent words are obtained, the speech synthesis model can be trained according to the plurality of first sample audios respectively corresponding to the plurality of first sample texts marked with the accent words and the plurality of first sample audios respectively corresponding to the plurality of first sample texts, and the plurality of second sample audios respectively corresponding to the second sample texts not marked with the accent words and the plurality of second sample texts.
In a possible manner, the speech synthesis model may comprise an accent recognition module, and accordingly the training step of the speech synthesis model may comprise: and firstly, generating an accent label corresponding to the second sample text through an accent recognition module. Then, sequentially using the first sample text and the second sample text as target sample texts, vectorizing a phoneme sequence corresponding to the target sample text to obtain a sample phoneme vector at a phoneme level, vectorizing a stress tag corresponding to the target sample text to obtain a sample stress tag vector, vectorizing a tone label corresponding to the target sample text to obtain a sample tone label vector, and determining the target sample phoneme vector according to the sample phoneme vector, the sample stress tag vector and the sample tone label vector. Then, a sample Mel frequency spectrum is determined according to the target sample phoneme vector, a first loss function is calculated according to the sample Mel frequency spectrum and the actual Mel frequency spectrum of the sample audio corresponding to the target sample text, and finally the parameters of the speech synthesis model are adjusted through the first loss function.
For example, the accent recognition module may be configured to recognize accent words in the second sample text that are not annotated with accent words and output corresponding accent labels. The accent labels may be represented by 0 and 1, for example, 0 may represent that no accent is labeled, and 1 represents that accent is labeled, so as to obtain accent labels represented by 0 and 1. For the training of the accent recognition module, joint training may be performed through the first sample text labeled with accent words in the process of training the speech synthesis model, or joint training may be performed through the sample text labeled with accent words different from the first sample text in the process of training the speech synthesis model, which is not limited in the embodiment of the present disclosure. It should be appreciated that, in order to reduce the amount of sample acquisition, it is preferable to jointly train the stress recognition module through the first sample labeled with the stress word.
For example, after the accent labels corresponding to the second sample text generated by the accent recognition module, the first sample text and the second sample text may be sequentially used as target sample texts, and the phoneme sequences corresponding to the target sample texts are vectorized to obtain sample phoneme vectors. It should be understood that phonemes are the smallest phonetic units divided according to the natural attributes of speech, and are divided into two broad categories, namely vowels and consonants. For Chinese, a phone includes an initial (the initial is a complete syllable formed by a consonant preceding a final and following the final) and a final (i.e., a vowel). For english, a phoneme includes a vowel and a consonant. In the embodiment of the present disclosure, a phoneme sequence corresponding to a target sample text is vectorized to obtain a sample phoneme vector, and a speech with phoneme-level stress can be synthesized in a subsequent process, so that the stress in the synthesized speech is controllable at a phoneme level, thereby further improving the accuracy of stress pronunciation in the synthesized speech. The process of vectorizing the phoneme sequence corresponding to the target sample text to obtain the sample phoneme vector is similar to the vector conversion method in the related art, and is not described herein again.
For example, if the target sample text is the first sample text, the accent label corresponding to the target sample text may be determined as follows: determining a phoneme sequence corresponding to a target sample text, and then performing accent labeling in the phoneme sequence corresponding to the target sample text according to accent words labeled in the target sample text, so as to obtain an accent label at a phoneme level corresponding to the target sample text. The stress labels may then be vectorized to obtain a sample stress label vector. If the target sample text is the second sample text, the accent label corresponding to the target sample text can be obtained through the accent recognition module. The method for obtaining the sample stress label vector by vectorizing the stress label is similar to the vector conversion method in the related art, and is not described herein again.
In the embodiment of the present disclosure, in order to synthesize a voice with a specific tone, in the training phase of the voice synthesis model, the tone label of the first speaker may be extracted from the first sample audio for model training, and simultaneously the tone label of the second speaker may be extracted from the second sample audio for model training. Specifically, after the first sample text and the second sample text are sequentially used as target sample texts, the tone labels corresponding to the target sample texts may be vectorized to obtain sample tone label vectors. For example, if the target sample text is the first sample text, the timbre label extracted from the first sample audio is vectorized to obtain a sample timbre label vector. And if the target sample text is a second sample text, vectorizing the tone label extracted from the second sample audio to obtain a sample tone label vector.
After the sample phoneme vector, the sample accent label vector and the sample tone label vector are obtained, three pieces of information which are mutually independent are considered to be represented by the sample phoneme vector, the sample accent label vector and the sample tone label vector, so that the target sample phoneme vector can be obtained by splicing the sample phoneme vector, the sample accent label vector and the sample tone label vector, but the target sample phoneme vector is not obtained by adding the sample phoneme vector, the sample accent label vector and the sample tone label vector, thereby avoiding damaging the content independence among the sample phoneme vector, the sample accent label vector and the sample tone label vector, and ensuring the accuracy of the output result of the speech synthesis model.
After the target sample phoneme vector is obtained, a sample Mel frequency spectrum may be determined from the target sample phoneme vector. In one possible approach, the target sample phoneme vector may be input to the encoder, and the vector output from the encoder may be input to the decoder to obtain a sample mel-frequency spectrum; the encoder is used for determining pronunciation information of each phoneme in a phoneme sequence corresponding to the input vector, and the decoder is used for performing conversion processing according to the pronunciation information of each phoneme corresponding to the input vector to obtain a Mel frequency spectrum corresponding to each phoneme. Or, the frame-level vector corresponding to the vector output by the encoder may be determined by an automatic alignment model, and then the frame-level vector is input to the decoder to obtain a sample mel spectrum, where the automatic alignment model is used to correspond the pronunciation information of the phoneme level in the sample text corresponding to the target sample phoneme vector to the frame time of each phoneme in the sample audio corresponding to the target sample phoneme vector, so as to improve the model training effect and further improve the accuracy of the accent pronunciation in the model synthesized speech.
For example, the speech synthesis model may be an end-to-end speech synthesis tacontron model, and accordingly, the encoder may be an encoder in the tacontron model and the decoder may be a decoder in the tacontron model. For example, as shown in fig. 2, in the training stage of the speech synthesis model, the phoneme sequence corresponding to the target sample text may be vectorized to obtain a vectorized phoneme sequence (i.e., a sample phoneme vector). Moreover, the voice synthesis model can divide the first sample text marked with the accent words into prosodic phrase sequences by level to obtain prosodic phrase sequences, then can generate word vector sequences according to the prosodic phrase sequences, and input the word vector sequences into the accent recognition module to train the accent recognition module. And then, vectorizing the stress label output by the stress identification module to obtain a vectorized stress label (i.e. a sample stress label vector). Meanwhile, the method can also carry out vectorization on the tone label corresponding to the sample audio corresponding to the target sample text to obtain a vectorized tone label (namely, the sample tone label).
Then, a target sample phoneme vector may be obtained by concatenating the sample phoneme vector, the sample accent label vector, and the sample tone label vector, and the target sample phoneme vector is input into an Encoder (Encoder) to obtain pronunciation information of each phoneme in the phoneme sequence corresponding to the target sample phoneme vector, for example, for a phoneme sequence corresponding to the target sample phoneme vector including a phoneme "jin", it is necessary to know that the pronunciation of the phoneme is the same as "today". Then, the alignment of phoneme level and frame level can be realized through an automatic alignment model, and a frame-level target sample vector corresponding to the vector output by the encoder is obtained. Then, the target sample vector may be input into a Decoder (Decoder) so that the Decoder performs a conversion process according to pronunciation information of each phoneme in a phoneme sequence corresponding to the target sample vector to obtain a sample Mel spectrum (Mel spectrum) corresponding to each phoneme.
In another possible approach, considering that the vector stitching process is set before the encoder, and the stress recognition module is calculated before the encoder, the processing speed is affected. Therefore, in order to increase the processing speed, referring to fig. 3, the concatenation process of the vectors may also be set after the encoder so that the accent recognition module and the encoder may perform parallel computations. That is, in the training stage of the speech synthesis model, the sample phoneme vector may be input into the encoder, and then the target sample phoneme vector may be determined according to the vector output by the encoder, the sample accent label vector, and the sample tone label vector. Accordingly, determining a sample mel-frequency spectrum from the target sample phoneme vector may be: and inputting the target sample phoneme vector into a decoder to obtain a sample Mel spectrum. By the method, the stress recognition module and the encoder can perform parallel calculation, so that the calculation speed of the speech synthesis model is increased, and the speech synthesis efficiency is improved.
After the sample mel spectrum is obtained, a first loss function may be calculated according to an actual mel spectrum of the sample audio corresponding to the sample mel spectrum and the target sample text, so that parameters of the speech synthesis model are adjusted by the first loss function. In a possible manner, the training step of the speech synthesis model further comprises: generating a word vector sequence corresponding to the first sample text, generating a sample stress label corresponding to the first sample text according to the stress word labeled by the first sample text, inputting the word vector sequence into a stress recognition module, and calculating a second loss function according to an output result of the stress recognition module and the sample stress label. Accordingly, adjusting parameters of the speech synthesis model by the first loss function may be: parameters of the speech synthesis model are adjusted by the first loss function and the second loss function.
For example, referring to fig. 4, the accent recognition module may take as input a word vector sequence corresponding to the first sample text and output as a predicted word-level accent label. The weighted second loss function may then be calculated from the accent labels output by the accent recognition module and the sample accent labels corresponding to the sample text. The word vector sequence may be obtained by performing word segmentation processing on a first sample text and then performing vector conversion on each word segmentation, and the sample stress label may be obtained by representing stress label information of each word segmentation by 0 and 1 according to a stress word labeled in the first sample text. The second Loss function may be a CE Loss function (Cross control Loss), etc., and the embodiment of the present disclosure does not limit this. It should be appreciated that, given the sparse nature of stress versus non-stress, stress tags may be given greater weight and non-stress tags may be given lesser weight in the calculation of the second loss function.
Considering that the target text is the second sample text, since the second sample text is not labeled with the accent word and the second loss function cannot be calculated, it may be determined whether the target text for training is labeled with the accent word first, and if the target text is labeled with the accent word, the parameters of the speech synthesis model are adjusted according to the calculated second loss function and the first loss function. If no accent words are labeled, parameters of the speech synthesis model may be adjusted by the first loss function. The first loss function may be an MSE loss function, and the second loss function may be a CE loss function, which is not limited in this disclosure. Or model optimization can be performed in the model training process through an Adam optimizer, so that the accuracy of the output result of the trained speech synthesis model is ensured.
In a possible approach, adjusting the parameters of the speech synthesis model by the first loss function and the second loss function may be: and carrying out weighted summation on the first loss function and the second loss function through the weight values which change in a self-adaptive manner in the training process to obtain a target loss function, and then adjusting the parameters of the voice synthesis model according to the target loss function.
For example, the weight value may be adaptively changed according to an order difference between the first loss function and the second loss function in each training process. For example, in the first training process, the calculation result of the first loss function is 10, the calculation result of the second loss function is 0.1, and the difference between the two is 2 orders of magnitude, so the weight value may be a value that makes the results of the first loss function and the second loss function be the same order of magnitude (for example, both are 1). In the second training process, the calculation result of the first loss function is 1, the second loss function is 0.1, and the difference between the two is 1 order of magnitude, so the weight value may be a numerical value that makes the first loss function unchanged and the result of the second loss function 1, and so on. Alternatively, the weight value may also be adaptively changed according to the degree of change of the values of the first loss function and the second loss function in each training process, and the like, which is not limited in this disclosure.
The first loss function and the second loss function can be subjected to weighted summation through the weight values which change in a self-adaptive mode in the training process, so that a target loss function is obtained, and therefore parameters of the voice synthesis model can be adjusted according to the target loss function. For example, in the above example, the CE loss function (i.e., the second loss function) and the MSE loss function (i.e., the first loss function) may be weighted and summed by using a weight value that adaptively changes during the training process to obtain a target loss function, and then parameters of the speech synthesis model are adjusted according to the target loss function, so as to implement the training of the speech synthesis model.
By the method, a speech synthesis model can be obtained according to the first sample audio of the first sample text marked with the accent words and the first sample audio of the first sample text by the first speaker, the second sample text not marked with the accent words and the second sample audio of the second sample text by the second speaker, and then speech synthesis can be performed on the text to be synthesized with the target tone label, wherein the target tone label is the tone label of the first speaker or the tone label of the second speaker.
Specifically, the speech synthesis model may determine an accent word in the text to be synthesized, and generate audio information corresponding to the text to be synthesized according to the accent word and the target tone label. The voice synthesis model can determine the accent words in the text to be synthesized through the trained accent recognition module. In other possible modes, the data difference between the training text of the accent recognition module and the text to be synthesized is considered, and the accuracy of the result output by the accent recognition module can be influenced, so that in order to improve the accuracy of accent pronunciation in the synthesized speech, the accent recognition function can be combined with a mode of manually marking accents. That is, the speech synthesis model determines that the accented words in the text to be synthesized may be: whether the text to be synthesized is marked with the accent words or not is determined, if the text to be synthesized is not marked with the accent words, the accent words in the text to be synthesized are identified through an accent identification module in a speech synthesis model, and if the text to be synthesized is marked with the accent words, the accent words in the text to be synthesized are determined according to marking information corresponding to the accent words.
By the method, the accent recognition module can be used for recognizing the accent words to synthesize accented voices under the condition that the accent words are not marked manually, and the accent words are synthesized with the accent voices according to the accent words marked manually under the condition that the accent words are marked manually, so that the requirement of voice synthesis under various scenes is met better, and the method is suitable for various different voice synthesis scenes.
After determining the accent words in the text to be synthesized, the audio information corresponding to the text to be synthesized may be generated according to the accent words and the target timbre label. In a possible mode, the speech synthesis model may first determine a phoneme sequence corresponding to the text to be synthesized, determine an accent label at a phoneme level according to the accent word, and then generate audio information corresponding to the text to be synthesized according to the phoneme sequence, the accent label, and the target timbre label corresponding to the text to be synthesized.
It should be understood that the speech synthesis model determines a phoneme sequence corresponding to a text to be synthesized, so that a speech with accents can be synthesized at a phoneme level in a subsequent process, so that the accents in the synthesized speech are controllable at the phoneme level, and the accuracy of accent pronunciation in the synthesized speech is further improved.
Simultaneously with or after determining the phoneme sequence corresponding to the text to be synthesized, the phoneme-level accent labels may be determined from the accent words. For example, the accent labels may be sequences represented as 0 and 1, where 0 represents that the corresponding phoneme is not marked with accent, and 1 represents that the corresponding phoneme in the text to be synthesized is marked with accent. In a specific application, a phoneme sequence corresponding to a text to be synthesized may be determined, and then, according to the accent word, accent labeling is performed in the phoneme sequence, so as to obtain an accent label at a phoneme level corresponding to the text to be synthesized.
After obtaining the phoneme sequence and the accent label corresponding to the text to be synthesized, the audio information corresponding to the text to be synthesized may be generated according to the phoneme sequence, the accent label, and the tone label carried by the text to be synthesized. In a possible mode, the speech synthesis model may first perform vectorization on a phoneme sequence corresponding to a text to be synthesized to obtain a phoneme vector, perform vectorization on an accent label to obtain an accent label vector, perform vectorization on a target tone label to obtain a tone label vector, then determine the target phoneme vector according to the phoneme vector, the accent label vector and the tone label vector, determine a mel spectrum according to the target phoneme vector, and finally input the mel spectrum into the vocoder to obtain audio information corresponding to the text to be synthesized.
It should be understood that the process of vectorizing the phoneme sequence corresponding to the text to be synthesized to obtain a phoneme vector, the process of vectorizing the accent label corresponding to the text to be synthesized to obtain an accent label vector, and the process of vectorizing the target timbre label to obtain a timbre label vector are similar to the vector conversion method in the related art, and are not described herein again.
In the embodiment of the present disclosure, it is considered that the phoneme vector, the accent label vector and the tone label vector represent two pieces of information that are independent of each other, and therefore, the target phoneme vector is obtained by splicing the phoneme vector, the accent label vector and the tone label vector, rather than obtaining the target phoneme vector by adding the phoneme vector, the accent label vector and the tone label vector, thereby avoiding destroying the content independence among the phoneme vector, the accent label vector and the tone label vector, and ensuring the accuracy of the subsequent speech synthesis result.
After the target phoneme vector is obtained, a Mel spectrum (Mel spectrum) may be determined from the target phoneme vector. In one possible approach, the target phoneme vector may be input into an encoder, and a vector result output by the encoder is input into a decoder to obtain a corresponding mel spectrum, where the encoder is configured to determine pronunciation information of each phoneme in a phoneme sequence corresponding to the input vector, and the decoder is configured to perform a conversion process according to the pronunciation information of each phoneme corresponding to the input vector to obtain the mel spectrum corresponding to each phoneme.
For example, referring to fig. 2, a speech synthesis model in an embodiment of the present disclosure may include an Encoder (Encoder) and a Decoder (Decoder). In the application stage of the speech synthesis model, after obtaining the target phoneme vector, the target phoneme vector may be input to the encoder to obtain pronunciation information of each phoneme in the phoneme sequence corresponding to the target phoneme vector, for example, for a phoneme "jin", it is necessary to know that the pronunciation of the phoneme is the same as "today". The pronunciation information may then be input into a decoder, so that the decoder performs a conversion process according to the pronunciation information of each phoneme in the phoneme sequence corresponding to the target phoneme vector to obtain a mel spectrum corresponding to each phoneme.
Alternatively, in other possible ways, to increase the processing speed, the concatenation process in the speech synthesis model may be arranged after the encoder, e.g. the speech synthesis model shown in fig. 3. In this case, in the application stage of the speech synthesis model, determining the target phoneme vector may be, according to the phoneme vector, the accent label vector, and the timbre label vector: the phoneme vector is input into an encoder, and then the target phoneme vector is determined according to a vector result output by the encoder, the accent label vector and the tone label vector. Accordingly, determining the mel spectrum from the target phoneme vector may be: the target phoneme vector is input to a decoder to obtain a Mel spectrum.
After determining the mel spectrum from the target phoneme vector, the mel spectrum may be input to the vocoder to obtain the audio information corresponding to the text to be synthesized. It should be understood that the embodiment of the present disclosure does not limit the type of the vocoder, that is, the audio information with accents can be obtained by inputting the mel spectrum into any vocoder, and the accents in the audio information can be obtained by the accent recognition module, so as to solve the problem in the related art that the synthesized speech has no accents or the accent pronunciation is incorrect due to randomly assigned accents, and improve the accuracy of the accent pronunciation in the synthesized speech. Further, the timbre of the synthesized speech may be controlled to synthesize accented speech having either the first speaker timbre or the second speaker timbre.
Based on the same inventive concept, the disclosed embodiments also provide a speech synthesis apparatus, which may become part or all of an electronic device through software, hardware, or a combination of both. Referring to fig. 5, the speech synthesis apparatus 500 includes:
the system comprises an acquisition module, a synthesis module and a processing module, wherein the acquisition module is used for acquiring a text to be synthesized with a target tone label, and the target tone label is a tone label of a first speaker or a tone label of a second speaker;
a synthesis module, configured to input the text to be synthesized into a speech synthesis model to obtain audio information corresponding to the text to be synthesized, where a training sample of the speech synthesis model includes a first sample text labeled with an accent word and a first sample audio of the first sample text by the first speaker, and a second sample text not labeled with an accent word and a second sample audio of the second sample text by the second speaker;
the voice synthesis model is used for determining accent words in the text to be synthesized and generating audio information corresponding to the text to be synthesized according to the accent words and the target tone label.
Optionally, the speech synthesis model generates audio information corresponding to the text to be synthesized through the following modules:
the first determining submodule is used for determining a phoneme sequence corresponding to the text to be synthesized;
the second determining submodule is used for determining the stress labels at the phoneme level according to the stress words;
and the generating submodule is used for generating audio information corresponding to the text to be synthesized according to the phoneme sequence corresponding to the text to be synthesized, the accent label and the target tone label.
Optionally, the generation submodule is configured to:
vectorizing a phoneme sequence corresponding to the text to be synthesized to obtain a phoneme vector;
vectorizing the stress label to obtain a stress label vector;
vectorizing the target tone label to obtain a tone label vector;
determining a target phoneme vector according to the phoneme vector, the accent label vector and the tone label vector;
determining a mel frequency spectrum according to the target phoneme vector;
and inputting the Mel frequency spectrum into a vocoder to obtain audio information corresponding to the text to be synthesized.
Optionally, the generation submodule is configured to:
and inputting the target phoneme vector into an encoder, and inputting the vector output by the encoder into a decoder to obtain a corresponding Mel frequency spectrum, wherein the encoder is used for determining pronunciation information of each phoneme in a phoneme sequence corresponding to the input vector, and the decoder is used for performing conversion processing according to the pronunciation information of each phoneme corresponding to the input vector to obtain the Mel frequency spectrum corresponding to each phoneme.
Optionally, the generation submodule is configured to:
inputting the phoneme vector into an encoder, and determining a target phoneme vector according to a vector output by the encoder, the accent label vector and the tone label vector;
inputting the target phoneme vector into a decoder to obtain the Mel frequency spectrum;
the encoder is configured to determine pronunciation information of each phoneme in a phoneme sequence corresponding to the input vector, and the decoder is configured to perform conversion processing according to the pronunciation information of each phoneme corresponding to the input vector to obtain a mel spectrum corresponding to each phoneme.
Optionally, the speech synthesis model comprises an accent recognition module, and the apparatus 500 further comprises the following modules for training the speech synthesis model:
the first sample generation module is used for generating an accent label corresponding to the second sample text through an accent recognition module of the speech synthesis model;
the first calculation module is used for sequentially taking the first sample text and the second sample text as target sample texts and vectorizing a phoneme sequence corresponding to the target sample texts to obtain a sample phoneme vector at a phoneme level; vectorizing the stress label corresponding to the target sample text to obtain a sample stress label vector; vectorizing a tone label corresponding to the target sample text to obtain a sample tone label vector; determining a target sample phoneme vector according to the sample phoneme vector, the sample stress label vector and the sample tone label vector; determining a sample Mel frequency spectrum according to the target sample phoneme vector; calculating a first loss function according to the sample Mel frequency spectrum and the actual Mel frequency spectrum of the sample audio corresponding to the target sample text;
and the adjusting module is used for adjusting the parameters of the speech synthesis model through the first loss function.
Optionally, the apparatus 500 further comprises the following modules for training the speech synthesis model:
a second sample generation module, configured to generate a word vector sequence corresponding to the first sample text, and generate a sample accent label corresponding to the first sample text according to the accent word labeled in the first sample text;
the second calculation module is used for inputting the word vector sequence into the stress recognition module and calculating a second loss function according to the output result of the stress recognition module and the sample stress label;
the adjusting module is used for adjusting the parameters of the speech synthesis model through the first loss function and the second loss function.
Optionally, the adjusting module is configured to:
weighting and summing the first loss function and the second loss function through a weight value which changes in a self-adaptive mode in the training process to obtain a target loss function;
and adjusting parameters of the speech synthesis model according to the target loss function.
Optionally, the speech synthesis model determines the accent words in the text to be synthesized by:
a third determining submodule, configured to determine whether a text to be synthesized is marked with a accent word;
the recognition sub-module is used for recognizing the accent words in the text to be synthesized through the accent recognition module in the speech synthesis model when the accent words are not marked in the text to be synthesized;
and the fourth determining submodule is used for determining the accent words in the text to be synthesized according to the marking information corresponding to the accent words when the accent words are marked in the text to be synthesized.
With regard to the apparatus in the above-described embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.
Based on the same inventive concept, the disclosed embodiments also provide a computer readable medium, on which a computer program is stored, which when executed by a processing device, implements the steps of any of the above-mentioned speech synthesis methods.
Based on the same inventive concept, an embodiment of the present disclosure further provides an electronic device, including:
a storage device having a computer program stored thereon;
processing means for executing the computer program in the storage means to implement the steps of any of the above-described speech synthesis methods.
Referring now to fig. 6, a schematic diagram of an electronic device (e.g., a terminal device or a server in fig. 1) 600 suitable for implementing embodiments of the present disclosure is shown. The terminal device in the embodiments of the present disclosure may include, but is not limited to, a mobile terminal such as a mobile phone, a notebook computer, a digital broadcast receiver, a PDA (personal digital assistant), a PAD (tablet computer), a PMP (portable multimedia player), a vehicle terminal (e.g., a car navigation terminal), and the like, and a stationary terminal such as a digital TV, a desktop computer, and the like. The electronic device shown in fig. 6 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present disclosure.
As shown in fig. 6, electronic device 600 may include a processing means (e.g., central processing unit, graphics processor, etc.) 601 that may perform various appropriate actions and processes in accordance with a program stored in a Read Only Memory (ROM)602 or a program loaded from a storage means 608 into a Random Access Memory (RAM) 603. In the RAM 603, various programs and data necessary for the operation of the electronic apparatus 600 are also stored. The processing device 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input/output (I/O) interface 605 is also connected to bus 604.
Generally, the following devices may be connected to the I/O interface 605: input devices 606 including, for example, a touch screen, touch pad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; output devices 607 including, for example, a Liquid Crystal Display (LCD), a speaker, a vibrator, and the like; storage 608 including, for example, tape, hard disk, etc.; and a communication device 609. The communication means 609 may allow the electronic device 600 to communicate with other devices wirelessly or by wire to exchange data. While fig. 6 illustrates an electronic device 600 having various means, it is to be understood that not all illustrated means are required to be implemented or provided. More or fewer devices may alternatively be implemented or provided.
In particular, according to an embodiment of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program carried on a non-transitory computer readable medium, the computer program containing program code for performing the method illustrated by the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network via the communication means 609, or may be installed from the storage means 608, or may be installed from the ROM 602. The computer program, when executed by the processing device 601, performs the above-described functions defined in the methods of the embodiments of the present disclosure.
It should be noted that the computer readable medium in the present disclosure can be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In contrast, in the present disclosure, a computer readable signal medium may comprise a propagated data signal with computer readable program code embodied therein, either in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, optical cables, RF (radio frequency), etc., or any suitable combination of the foregoing.
In some embodiments, the communication may be performed using any currently known or future developed network Protocol, such as HTTP (HyperText Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), the Internet (e.g., the Internet), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future developed network.
The computer readable medium may be embodied in the electronic device; or may exist separately without being assembled into the electronic device.
The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to: acquiring a text to be synthesized with a target tone label, wherein the target tone label is the tone label of the first speaker or the tone label of the second speaker; inputting the text to be synthesized into a speech synthesis model to obtain audio information corresponding to the text to be synthesized, wherein training samples of the speech synthesis model comprise a first sample text marked with accent words, a first sample audio of a first speaker to the first sample text, and a second sample text not marked with accent words and a second sample audio of a second speaker to the second sample text; the voice synthesis model is used for determining accent words in the text to be synthesized and generating audio information corresponding to the text to be synthesized according to the accent words and the target tone label.
Computer program code for carrying out operations for the present disclosure may be written in any combination of one or more programming languages, including but not limited to an object oriented programming language such as Java, Smalltalk, C + +, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The modules described in the embodiments of the present disclosure may be implemented by software or hardware. Wherein the name of a module in some cases does not constitute a limitation on the module itself.
The functions described herein above may be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), Application Specific Standard Products (ASSPs), systems on a chip (SOCs), Complex Programmable Logic Devices (CPLDs), and the like.
In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
Example 1 provides a speech synthesis method according to one or more embodiments of the present disclosure, the method including:
acquiring a text to be synthesized with a target tone label, wherein the target tone label is a tone label of a first speaker or a tone label of a second speaker;
inputting the text to be synthesized into a speech synthesis model to obtain audio information corresponding to the text to be synthesized, wherein training samples of the speech synthesis model comprise a first sample text marked with accent words, a first sample audio of the first speaker to the first sample text, a second sample text not marked with accent words, and a second sample audio of the second speaker to the second sample text;
the voice synthesis model is used for determining accent words in the text to be synthesized and generating audio information corresponding to the text to be synthesized according to the accent words and the target tone label.
Example 2 provides the method of example 1, the speech synthesis model generating audio information corresponding to the text to be synthesized by:
determining a phoneme sequence corresponding to the text to be synthesized;
determining an accent label of a phoneme level according to the accent word;
and generating audio information corresponding to the text to be synthesized according to the phoneme sequence corresponding to the text to be synthesized, the accent label and the target tone label.
Example 3 provides the method of example 2, wherein generating audio information corresponding to the text to be synthesized according to the phoneme sequence, the accent label, and the target timbre label includes:
vectorizing the phoneme sequence to obtain a phoneme vector;
vectorizing the stress label to obtain a stress label vector;
vectorizing the target tone label to obtain a tone label vector;
determining a target phoneme vector according to the phoneme vector, the accent label vector and the tone label vector;
determining a mel frequency spectrum according to the target phoneme vector;
and inputting the Mel frequency spectrum into a vocoder to obtain audio information corresponding to the text to be synthesized.
Example 4 provides the method of example 3, determining a mel spectrum from the target phoneme vector, comprising:
and inputting the target phoneme vector into an encoder, and inputting the vector output by the encoder into a decoder to obtain a corresponding Mel frequency spectrum, wherein the encoder is used for determining pronunciation information of each phoneme in a phoneme sequence corresponding to the input vector, and the decoder is used for performing conversion processing according to the pronunciation information of each phoneme corresponding to the input vector to obtain the Mel frequency spectrum corresponding to each phoneme.
Example 5 provides the method of example 3, wherein determining the target phoneme vector from the phoneme vector, the accent label vector, and the timbre label vector comprises:
inputting the phoneme vector into an encoder, and determining the target phoneme vector according to the vector output by the encoder, the accent label vector and the tone label vector;
the determining a mel frequency spectrum according to the target phoneme vector comprises:
inputting the target phoneme vector into a decoder to obtain the Mel frequency spectrum;
the encoder is configured to determine pronunciation information of each phoneme in a phoneme sequence corresponding to the input vector, and the decoder is configured to perform conversion processing according to the pronunciation information of each phoneme corresponding to the input vector to obtain a mel spectrum corresponding to each phoneme.
Example 6 provides the method of any one of examples 1 to 5, the speech synthesis model including an accent recognition module, the training of the speech synthesis model including:
generating an accent label corresponding to the second sample text through the accent recognition module;
sequentially taking the first sample text and the second sample text as target sample texts, and vectorizing a phoneme sequence corresponding to the target sample text to obtain a sample phoneme vector at a phoneme level; vectorizing the stress label corresponding to the target sample text to obtain a sample stress label vector; vectorizing a tone label corresponding to the target sample text to obtain a sample tone label vector; determining a target sample phoneme vector according to the sample phoneme vector, the sample stress label vector and the sample tone label vector; determining a sample Mel frequency spectrum according to the target sample phoneme vector; calculating a first loss function according to the sample Mel frequency spectrum and the actual Mel frequency spectrum of the sample audio corresponding to the target sample text;
parameters of the speech synthesis model are adjusted by the first loss function.
Example 7 provides the method of example 6, the training step further comprising, in accordance with one or more embodiments of the present disclosure:
generating a word vector sequence corresponding to the first sample text, and generating a sample accent label corresponding to the first sample text according to the accent words labeled by the first sample text;
inputting the word vector sequence into the stress recognition module, and calculating a second loss function according to an output result of the stress recognition module and the sample stress label;
the adjusting parameters of the speech synthesis model by the first loss function includes:
adjusting parameters of the speech synthesis model by the first loss function and the second loss function.
In accordance with one or more embodiments of the present disclosure, example 8 provides the method of example 7,
the adjusting parameters of the speech synthesis model by the first loss function and the second loss function comprises:
weighting and summing the first loss function and the second loss function through a weight value which changes in a self-adaptive mode in the training process to obtain a target loss function;
and adjusting parameters of the speech synthesis model according to the target loss function.
Example 9 provides the method of example 6, the determining the accent words in the text to be synthesized, including:
determining whether the text to be synthesized is marked with accent words;
if the text to be synthesized is not marked with the accent words, identifying the accent words in the text to be synthesized through the accent identification module;
and if the text to be synthesized is marked with the accent words, determining the accent words in the text to be synthesized according to the marking information corresponding to the accent words.
Example 10 provides, in accordance with one or more embodiments of the present disclosure, a speech synthesis apparatus, the apparatus comprising:
the system comprises an acquisition module, a synthesis module and a processing module, wherein the acquisition module is used for acquiring a text to be synthesized with a target tone label, and the target tone label is a tone label of a first speaker or a tone label of a second speaker;
the synthesis module is used for inputting the text to be synthesized into a speech synthesis model so as to obtain audio information corresponding to the text to be synthesized;
wherein the training samples of the speech synthesis model comprise a first sample text marked with accent words and a first sample audio of the first speaker to the first sample text, and a second sample text not marked with accent words and a second sample audio of the second speaker to the second sample text;
the voice synthesis model is used for determining accent words in the text to be synthesized and generating audio information corresponding to the text to be synthesized according to the accent words and the target tone label.
Example 11 provides the apparatus of example 10, the speech synthesis model to generate audio information corresponding to the text to be synthesized by:
the first determining submodule is used for determining a phoneme sequence corresponding to the text to be synthesized;
the second determining submodule is used for determining the stress labels at the phoneme level according to the stress words;
and the generating submodule is used for generating audio information corresponding to the text to be synthesized according to the phoneme sequence corresponding to the text to be synthesized, the accent label and the target tone label.
Example 12 provides the apparatus of example 11, the generation submodule to:
vectorizing a phoneme sequence corresponding to the text to be synthesized to obtain a phoneme vector;
vectorizing the stress label to obtain a stress label vector;
vectorizing the target tone label to obtain a tone label vector;
determining a target phoneme vector according to the phoneme vector, the accent label vector and the tone label vector;
determining a mel frequency spectrum according to the target phoneme vector;
and inputting the Mel frequency spectrum into a vocoder to obtain audio information corresponding to the text to be synthesized.
Example 13 provides the apparatus of example 12, the generation submodule to:
and inputting the target phoneme vector into an encoder, and inputting the vector output by the encoder into a decoder to obtain a corresponding Mel frequency spectrum, wherein the encoder is used for determining pronunciation information of each phoneme in a phoneme sequence corresponding to the input vector, and the decoder is used for performing conversion processing according to the pronunciation information of each phoneme corresponding to the input vector to obtain the Mel frequency spectrum corresponding to each phoneme.
Example 14 provides the apparatus of example 12, the generation submodule to:
inputting the phoneme vector into an encoder, and determining a target phoneme vector according to a vector output by the encoder, the accent label vector and the tone label vector;
inputting the target phoneme vector into a decoder to obtain the Mel frequency spectrum;
the encoder is configured to determine pronunciation information of each phoneme in a phoneme sequence corresponding to the input vector, and the decoder is configured to perform conversion processing according to the pronunciation information of each phoneme corresponding to the input vector to obtain a mel spectrum corresponding to each phoneme.
Example 15 provides the apparatus of any one of examples 10 to 14, the speech synthesis model including an accent recognition module, the apparatus 500 further including the following modules for training the speech synthesis model:
the first sample generation module is used for generating an accent label corresponding to the second sample text through an accent recognition module of the speech synthesis model;
the first calculation module is used for sequentially taking the first sample text and the second sample text as target sample texts and vectorizing a phoneme sequence corresponding to the target sample texts to obtain a sample phoneme vector at a phoneme level; vectorizing the stress label corresponding to the target sample text to obtain a sample stress label vector; vectorizing a tone label corresponding to the target sample text to obtain a sample tone label vector; determining a target sample phoneme vector according to the sample phoneme vector, the sample stress label vector and the sample tone label vector; determining a sample Mel frequency spectrum according to the target sample phoneme vector; calculating a first loss function according to the sample Mel frequency spectrum and the actual Mel frequency spectrum of the sample audio corresponding to the target sample text;
and the adjusting module is used for adjusting the parameters of the speech synthesis model through the first loss function.
Example 16 provides the apparatus of example 15, further comprising the following modules for training the speech synthesis model, in accordance with one or more embodiments of the present disclosure:
a second sample generation module, configured to generate a word vector sequence corresponding to the first sample text, and generate a sample accent label corresponding to the first sample text according to the accent word labeled in the first sample text;
the second calculation module is used for inputting the word vector sequence into the stress recognition module and calculating a second loss function according to the output result of the stress recognition module and the sample stress label;
the adjusting module is used for adjusting the parameters of the speech synthesis model through the first loss function and the second loss function.
Example 17 provides the apparatus of example 16, the adjustment module to:
weighting and summing the first loss function and the second loss function through a weight value which changes in a self-adaptive mode in the training process to obtain a target loss function;
and adjusting parameters of the speech synthesis model according to the target loss function.
Example 18 provides the apparatus of example 15, the speech synthesis model to determine accented words in the text to be synthesized by:
a third determining submodule, configured to determine whether a text to be synthesized is marked with a accent word;
the recognition sub-module is used for recognizing the accent words in the text to be synthesized through the accent recognition module in the speech synthesis model when the accent words are not marked in the text to be synthesized;
and the fourth determining submodule is used for determining the accent words in the text to be synthesized according to the marking information corresponding to the accent words when the accent words are marked in the text to be synthesized.
Example 19 provides a computer-readable medium on which is stored a computer program that, when executed by a processing apparatus, implements the steps of the speech synthesis method of any one of examples 1 to 9.
Example 20 provides, in accordance with one or more embodiments of the present disclosure, an electronic device comprising:
a storage device having a computer program stored thereon;
processing means for executing the computer program in the storage means to implement the steps of the speech synthesis method in any one of examples 1 to 9.
The foregoing description is only exemplary of the preferred embodiments of the disclosure and is illustrative of the principles of the technology employed. It will be appreciated by those skilled in the art that the scope of the disclosure herein is not limited to the particular combination of features described above, but also encompasses other embodiments in which any combination of the features described above or their equivalents does not depart from the spirit of the disclosure. For example, the above features and (but not limited to) the features disclosed in this disclosure having similar functions are replaced with each other to form the technical solution.
Further, while operations are depicted in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Likewise, while several specific implementation details are included in the above discussion, these should not be construed as limitations on the scope of the disclosure. Certain features that are described in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable subcombination.
Although the subject matter has been described in language specific to structural features and/or methodological acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed as example forms of implementing the claims. With regard to the apparatus in the above-described embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.
Claims (12)
1. A method of speech synthesis, the method comprising:
acquiring a text to be synthesized with a target tone label, wherein the target tone label is a tone label of a first speaker or a tone label of a second speaker;
inputting the text to be synthesized into a speech synthesis model to obtain audio information corresponding to the text to be synthesized, wherein training samples of the speech synthesis model comprise a first sample text marked with accent words, a first sample audio of the first speaker to the first sample text, a second sample text not marked with accent words, and a second sample audio of the second speaker to the second sample text;
the voice synthesis model is used for determining accent words in the text to be synthesized and generating audio information corresponding to the text to be synthesized according to the accent words and the target tone label.
2. The method of claim 1, wherein the speech synthesis model generates the audio information corresponding to the text to be synthesized by:
determining a phoneme sequence corresponding to the text to be synthesized;
determining an accent label of a phoneme level according to the accent word;
and generating audio information corresponding to the text to be synthesized according to the phoneme sequence, the accent label and the target tone label.
3. The method according to claim 2, wherein the generating audio information corresponding to the text to be synthesized according to the phoneme sequence, the accent label and the target timbre label comprises:
vectorizing the phoneme sequence to obtain a phoneme vector;
vectorizing the stress label to obtain a stress label vector;
vectorizing the target tone label to obtain a tone label vector;
determining a target phoneme vector according to the phoneme vector, the accent label vector and the tone label vector;
determining a mel frequency spectrum according to the target phoneme vector;
and inputting the Mel frequency spectrum into a vocoder to obtain audio information corresponding to the text to be synthesized.
4. The method of claim 3, wherein determining a Mel frequency spectrum from the target phoneme vector comprises:
and inputting the target phoneme vector into an encoder, and inputting the vector output by the encoder into a decoder to obtain a corresponding Mel frequency spectrum, wherein the encoder is used for determining pronunciation information of each phoneme in a phoneme sequence corresponding to the input vector, and the decoder is used for performing conversion processing according to the pronunciation information of each phoneme corresponding to the input vector to obtain the Mel frequency spectrum corresponding to each phoneme.
5. The method of claim 3, wherein determining a target phoneme vector based on the phoneme vector, the accent label vector, and the timbre label vector comprises:
inputting the phoneme vector into an encoder, and determining the target phoneme vector according to the vector output by the encoder, the accent label vector and the tone label vector;
the determining a mel frequency spectrum according to the target phoneme vector comprises:
inputting the target phoneme vector into a decoder to obtain the Mel frequency spectrum;
the encoder is configured to determine pronunciation information of each phoneme in a phoneme sequence corresponding to the input vector, and the decoder is configured to perform conversion processing according to the pronunciation information of each phoneme corresponding to the input vector to obtain a mel spectrum corresponding to each phoneme.
6. The method according to any of claims 1-5, wherein the speech synthesis model comprises an accent recognition module, and wherein the training of the speech synthesis model comprises:
generating an accent label corresponding to the second sample text through the accent recognition module;
sequentially taking the first sample text and the second sample text as target sample texts, and vectorizing a phoneme sequence corresponding to the target sample text to obtain a sample phoneme vector at a phoneme level; vectorizing the stress label corresponding to the target sample text to obtain a sample stress label vector; vectorizing a tone label corresponding to the target sample text to obtain a sample tone label vector; determining a target sample phoneme vector according to the sample phoneme vector, the sample stress label vector and the sample tone label vector; determining a sample Mel frequency spectrum according to the target sample phoneme vector; calculating a first loss function according to the sample Mel frequency spectrum and the actual Mel frequency spectrum of the sample audio corresponding to the target sample text;
parameters of the speech synthesis model are adjusted by the first loss function.
7. The method of claim 6, wherein the training step further comprises:
generating a word vector sequence corresponding to the first sample text, and generating a sample accent label corresponding to the first sample text according to the accent words labeled by the first sample text;
inputting the word vector sequence into the stress recognition module, and calculating a second loss function according to an output result of the stress recognition module and the sample stress label;
the adjusting parameters of the speech synthesis model by the first loss function includes:
adjusting parameters of the speech synthesis model by the first loss function and the second loss function.
8. The method of claim 7, wherein the adjusting parameters of the speech synthesis model by the first loss function and the second loss function comprises:
weighting and summing the first loss function and the second loss function through a weight value which changes in a self-adaptive mode in the training process to obtain a target loss function;
and adjusting parameters of the speech synthesis model according to the target loss function.
9. The method of claim 6, wherein the determining the accent words in the text to be synthesized comprises:
determining whether the text to be synthesized is marked with accent words;
if the text to be synthesized is not marked with the accent words, identifying the accent words in the text to be synthesized through the accent identification module;
and if the text to be synthesized is marked with the accent words, determining the accent words in the text to be synthesized according to the marking information corresponding to the accent words.
10. A speech synthesis apparatus, characterized in that the apparatus comprises:
the system comprises an acquisition module, a synthesis module and a processing module, wherein the acquisition module is used for acquiring a text to be synthesized with a target tone label, and the target tone label is a tone label of a first speaker or a tone label of a second speaker;
the synthesis module is used for inputting the text to be synthesized into a speech synthesis model so as to obtain audio information corresponding to the text to be synthesized;
wherein the training samples of the speech synthesis model comprise a first sample text marked with accent words and a first sample audio of the first speaker to the first sample text, and a second sample text not marked with accent words and a second sample audio of the second speaker to the second sample text;
the voice synthesis model is used for determining accent words in the text to be synthesized and generating audio information corresponding to the text to be synthesized according to the accent words and the target tone label.
11. A computer-readable medium, on which a computer program is stored, characterized in that the program, when being executed by processing means, carries out the steps of the method of any one of claims 1-9.
12. An electronic device, comprising:
a storage device having a computer program stored thereon;
processing means for executing the computer program in the storage means to carry out the steps of the method according to any one of claims 1 to 9.
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011211084.5A CN112309366B (en) | 2020-11-03 | 2020-11-03 | Speech synthesis method, speech synthesis device, storage medium and electronic equipment |
US18/042,077 US20230317055A1 (en) | 2020-11-03 | 2021-10-25 | Method, apparatus, storage medium and electronic device for speech synthesis |
PCT/CN2021/126138 WO2022095743A1 (en) | 2020-11-03 | 2021-10-25 | Speech synthesis method and apparatus, storage medium, and electronic device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011211084.5A CN112309366B (en) | 2020-11-03 | 2020-11-03 | Speech synthesis method, speech synthesis device, storage medium and electronic equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112309366A true CN112309366A (en) | 2021-02-02 |
CN112309366B CN112309366B (en) | 2022-06-14 |
Family
ID=74332713
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011211084.5A Active CN112309366B (en) | 2020-11-03 | 2020-11-03 | Speech synthesis method, speech synthesis device, storage medium and electronic equipment |
Country Status (3)
Country | Link |
---|---|
US (1) | US20230317055A1 (en) |
CN (1) | CN112309366B (en) |
WO (1) | WO2022095743A1 (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113035169A (en) * | 2021-03-12 | 2021-06-25 | 北京帝派智能科技有限公司 | Voice synthesis method and system capable of training personalized tone library on line |
CN113096634A (en) * | 2021-03-30 | 2021-07-09 | 平安科技(深圳)有限公司 | Speech synthesis method, apparatus, server and storage medium |
CN113270090A (en) * | 2021-05-19 | 2021-08-17 | 平安科技(深圳)有限公司 | Combined model training method and device based on ASR model and TTS model |
CN113409761A (en) * | 2021-07-12 | 2021-09-17 | 上海喜马拉雅科技有限公司 | Speech synthesis method, speech synthesis device, electronic equipment and computer-readable storage medium |
CN113450759A (en) * | 2021-06-22 | 2021-09-28 | 北京百度网讯科技有限公司 | Voice generation method, device, electronic equipment and storage medium |
CN113658577A (en) * | 2021-08-16 | 2021-11-16 | 腾讯音乐娱乐科技(深圳)有限公司 | Speech synthesis model training method, audio generation method, device and medium |
WO2022095743A1 (en) * | 2020-11-03 | 2022-05-12 | 北京有竹居网络技术有限公司 | Speech synthesis method and apparatus, storage medium, and electronic device |
CN114613353A (en) * | 2022-03-25 | 2022-06-10 | 马上消费金融股份有限公司 | Speech synthesis method, speech synthesis device, electronic equipment and storage medium |
CN114863946A (en) * | 2022-04-26 | 2022-08-05 | 华南理工大学 | Voice cloning method based on phoneme level speaker migration |
WO2023206928A1 (en) * | 2022-04-27 | 2023-11-02 | 网易(杭州)网络有限公司 | Speech processing method and apparatus, computer device, and computer-readable storage medium |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117672182B (en) * | 2024-02-02 | 2024-06-07 | 江西拓世智能科技股份有限公司 | Sound cloning method and system based on artificial intelligence |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2008046636A (en) * | 2006-08-14 | 2008-02-28 | Pentax Corp | Japanese speech synthesizing method and system using accent phrase matching prior select |
CN101359473A (en) * | 2007-07-30 | 2009-02-04 | 国际商业机器公司 | Auto speech conversion method and apparatus |
US20110202345A1 (en) * | 2010-02-12 | 2011-08-18 | Nuance Communications, Inc. | Method and apparatus for generating synthetic speech with contrastive stress |
CN102254554A (en) * | 2011-07-18 | 2011-11-23 | 中国科学院自动化研究所 | Method for carrying out hierarchical modeling and predicating on mandarin accent |
CN105304080A (en) * | 2015-09-22 | 2016-02-03 | 科大讯飞股份有限公司 | Speech synthesis device and speech synthesis method |
US20200082807A1 (en) * | 2018-01-11 | 2020-03-12 | Neosapience, Inc. | Text-to-speech synthesis method and apparatus using machine learning, and computer-readable storage medium |
CN111667810A (en) * | 2020-06-08 | 2020-09-15 | 北京有竹居网络技术有限公司 | Method and device for acquiring polyphone corpus, readable medium and electronic equipment |
Family Cites Families (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107705783B (en) * | 2017-11-27 | 2022-04-26 | 北京搜狗科技发展有限公司 | Voice synthesis method and device |
US10418025B2 (en) * | 2017-12-06 | 2019-09-17 | International Business Machines Corporation | System and method for generating expressive prosody for speech synthesis |
WO2020027619A1 (en) * | 2018-08-02 | 2020-02-06 | 네오사피엔스 주식회사 | Method, device, and computer readable storage medium for text-to-speech synthesis using machine learning on basis of sequential prosody feature |
CN110148398A (en) * | 2019-05-16 | 2019-08-20 | 平安科技(深圳)有限公司 | Training method, device, equipment and the storage medium of speech synthesis model |
CN110264992B (en) * | 2019-06-11 | 2021-03-16 | 百度在线网络技术(北京)有限公司 | Speech synthesis processing method, apparatus, device and storage medium |
CN111161702B (en) * | 2019-12-23 | 2022-08-26 | 爱驰汽车有限公司 | Personalized speech synthesis method and device, electronic equipment and storage medium |
CN112309366B (en) * | 2020-11-03 | 2022-06-14 | 北京有竹居网络技术有限公司 | Speech synthesis method, speech synthesis device, storage medium and electronic equipment |
-
2020
- 2020-11-03 CN CN202011211084.5A patent/CN112309366B/en active Active
-
2021
- 2021-10-25 WO PCT/CN2021/126138 patent/WO2022095743A1/en active Application Filing
- 2021-10-25 US US18/042,077 patent/US20230317055A1/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2008046636A (en) * | 2006-08-14 | 2008-02-28 | Pentax Corp | Japanese speech synthesizing method and system using accent phrase matching prior select |
CN101359473A (en) * | 2007-07-30 | 2009-02-04 | 国际商业机器公司 | Auto speech conversion method and apparatus |
US20110202345A1 (en) * | 2010-02-12 | 2011-08-18 | Nuance Communications, Inc. | Method and apparatus for generating synthetic speech with contrastive stress |
CN102254554A (en) * | 2011-07-18 | 2011-11-23 | 中国科学院自动化研究所 | Method for carrying out hierarchical modeling and predicating on mandarin accent |
CN105304080A (en) * | 2015-09-22 | 2016-02-03 | 科大讯飞股份有限公司 | Speech synthesis device and speech synthesis method |
US20200082807A1 (en) * | 2018-01-11 | 2020-03-12 | Neosapience, Inc. | Text-to-speech synthesis method and apparatus using machine learning, and computer-readable storage medium |
CN111667810A (en) * | 2020-06-08 | 2020-09-15 | 北京有竹居网络技术有限公司 | Method and device for acquiring polyphone corpus, readable medium and electronic equipment |
Non-Patent Citations (2)
Title |
---|
N.MINEMATSU,ET AL.: "Automatic estimation of accentual attribute values of words to realize accent sandhi in Japanese text-to-speech conversion", 《PROCEEDINGS OF 2002 IEEE WORKSHOP ON SPEECH SYNTHESIS, 2002.》 * |
孟凡博等: "汉语重音的凸显度分析与合成", 《声学学报》 * |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2022095743A1 (en) * | 2020-11-03 | 2022-05-12 | 北京有竹居网络技术有限公司 | Speech synthesis method and apparatus, storage medium, and electronic device |
CN113035169A (en) * | 2021-03-12 | 2021-06-25 | 北京帝派智能科技有限公司 | Voice synthesis method and system capable of training personalized tone library on line |
CN113035169B (en) * | 2021-03-12 | 2021-12-07 | 北京帝派智能科技有限公司 | Voice synthesis method and system capable of training personalized tone library on line |
CN113096634A (en) * | 2021-03-30 | 2021-07-09 | 平安科技(深圳)有限公司 | Speech synthesis method, apparatus, server and storage medium |
CN113096634B (en) * | 2021-03-30 | 2024-03-01 | 平安科技(深圳)有限公司 | Speech synthesis method, device, server and storage medium |
CN113270090A (en) * | 2021-05-19 | 2021-08-17 | 平安科技(深圳)有限公司 | Combined model training method and device based on ASR model and TTS model |
CN113450759A (en) * | 2021-06-22 | 2021-09-28 | 北京百度网讯科技有限公司 | Voice generation method, device, electronic equipment and storage medium |
CN113409761A (en) * | 2021-07-12 | 2021-09-17 | 上海喜马拉雅科技有限公司 | Speech synthesis method, speech synthesis device, electronic equipment and computer-readable storage medium |
CN113658577A (en) * | 2021-08-16 | 2021-11-16 | 腾讯音乐娱乐科技(深圳)有限公司 | Speech synthesis model training method, audio generation method, device and medium |
CN114613353A (en) * | 2022-03-25 | 2022-06-10 | 马上消费金融股份有限公司 | Speech synthesis method, speech synthesis device, electronic equipment and storage medium |
CN114613353B (en) * | 2022-03-25 | 2023-08-08 | 马上消费金融股份有限公司 | Speech synthesis method, device, electronic equipment and storage medium |
CN114863946A (en) * | 2022-04-26 | 2022-08-05 | 华南理工大学 | Voice cloning method based on phoneme level speaker migration |
WO2023206928A1 (en) * | 2022-04-27 | 2023-11-02 | 网易(杭州)网络有限公司 | Speech processing method and apparatus, computer device, and computer-readable storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN112309366B (en) | 2022-06-14 |
US20230317055A1 (en) | 2023-10-05 |
WO2022095743A1 (en) | 2022-05-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112309366B (en) | Speech synthesis method, speech synthesis device, storage medium and electronic equipment | |
CN111292720B (en) | Speech synthesis method, device, computer readable medium and electronic equipment | |
CN112331176B (en) | Speech synthesis method, speech synthesis device, storage medium and electronic equipment | |
CN111369967B (en) | Virtual character-based voice synthesis method, device, medium and equipment | |
CN111369971B (en) | Speech synthesis method, device, storage medium and electronic equipment | |
US11450313B2 (en) | Determining phonetic relationships | |
CN112786011B (en) | Speech synthesis method, synthesis model training method, device, medium and equipment | |
CN112309367B (en) | Speech synthesis method, speech synthesis device, storage medium and electronic equipment | |
KR20210103002A (en) | Speech synthesis method and apparatus based on emotion information | |
CN111192568B (en) | Speech synthesis method and speech synthesis device | |
CN111583900A (en) | Song synthesis method and device, readable medium and electronic equipment | |
CN112786007A (en) | Speech synthesis method, device, readable medium and electronic equipment | |
KR20160058470A (en) | Speech synthesis apparatus and control method thereof | |
CN112802446B (en) | Audio synthesis method and device, electronic equipment and computer readable storage medium | |
CN113658577B (en) | Speech synthesis model training method, audio generation method, equipment and medium | |
WO2023160553A1 (en) | Speech synthesis method and apparatus, and computer-readable medium and electronic device | |
CN113327580A (en) | Speech synthesis method, device, readable medium and electronic equipment | |
CN113808571B (en) | Speech synthesis method, speech synthesis device, electronic device and storage medium | |
CN113421550A (en) | Speech synthesis method, device, readable medium and electronic equipment | |
WO2008147649A1 (en) | Method for synthesizing speech | |
CN114255738A (en) | Speech synthesis method, apparatus, medium, and electronic device | |
CN115101046A (en) | Method and device for synthesizing voice of specific speaker | |
KR20080049813A (en) | Speech dialog method and device | |
CN114464164A (en) | Speech synthesis method, device, readable medium and electronic equipment | |
CN114242035A (en) | Speech synthesis method, apparatus, medium, and electronic device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |